void SoftReset() {
BYTE ipl_backup = SRbits.IPL;
SRbits.IPL = 7; // disable interrupts
log_printf("SoftReset()");
ucReset(); // stop uJVM
TimersInit();
PinsInit();
PWMInit();
ADCInit();
UARTInit();
SPIInit();
I2CInit();
InCapInit();
ucInit(); // restart uJVM
// TODO: reset all peripherals!
SRbits.IPL = ipl_backup; // enable interrupts
}
int main(int argc, char *argv[])
{
char *encoding = NULL;
bool doHash = false;
if (argc < 2){
fprintf(stderr, "Usage: %s [ -e encoding ] filename [ [ -e encoding] filename2 ] ...\n", argv[0]);
exit(1);
}
ucInit();
if ( ! hashinit() ) {
log("db: Failed to init hashtable." ); return 1; }
// . hashinit() calls srand() w/ a fixed number
// . let's mix it up again
srand ( time(NULL) );
int i;
int flag = flNone;
void (*parser)(char*,int,bool,char*) = NULL;
char *parser_str = "";
for (i=1;i<argc;i++){
// Read cmdline args
if (argv[i][0] == '-'){
if (gbstrlen(argv[i])<2){
fprintf(stderr, "Unknown argument: %s\n",
argv[i]);
exit(1);
}
switch(argv[i][1]){
case 'e':
flag = flEncoding;
break;
case 'p':
flag = flParser;
break;
case 'h':
flag = flHash;
break;
case 's':
flag = flFilterSpaces;
break;
default:
fprintf(stderr, "Unknown flag: %s\n",
argv[i]);
exit(1);
}
continue; //next arg
}
// Switch default encoding
if (flag == flEncoding){
encoding = argv[i];
flag=flNone;
fprintf(stderr, "Using encoding: %s\n", encoding);
continue;
}
// switch parser
if (flag == flParser){
if (!strncmp(argv[i], "icu", 3)){
parser = parse_doc_icu;
parser_str = "ICU BreakIterator";
}
else {
parser = parse_doc_8859_1;
parser_str = "iso-8859-1";
}
flag=flNone;
fprintf(stderr, "Using parser: %s\n", parser_str);
continue;
}
if (flag == flHash){
if ((!strncmp(argv[i], "0", 1)) ||
(!strncmp(argv[i], "f", 1))){
doHash = false;
}
else{ doHash = true; }
flag = flNone;
continue;
}
if (flag == flFilterSpaces){
if ((!strncmp(argv[i], "0", 1)) ||
(!strncmp(argv[i], "f", 1))){
doFilterSpaces = false;
}
else{ doFilterSpaces = true; }
flag = flNone;
continue;
}
char * filename = argv[i];
fprintf(stderr, "Reading \"%s\"\n", filename);
FILE *fp = fopen(filename,"r");
if (!fp){
fprintf(stderr, "Error: could not open file \"%s\"\n",
filename);
continue;
}
// Get File size
size_t file_size;
fseek(fp, 0L, SEEK_END);
file_size = (size_t)ftell(fp);
fseek(fp, 0L, SEEK_SET);
char *file_buf = (char*)malloc(file_size+1);
size_t nread = fread(file_buf, (size_t)1,file_size, fp);
fclose(fp);
if (nread != file_size){
fprintf(stderr, "Warning: wanted %d chars, but read %d\n",
file_size, nread);
}
file_buf[nread] = '\0';
//struct timeval tv1, tv2;
//struct timezone tz1, tz2;
long usec_elapsed;
// int testnum;
char ucBuf[128*1024];
long ucLen = ucToUnicode((UChar*)ucBuf,128*1024,
file_buf,nread+1,"utf-8", 10);
ucLen <<= 1;
usec_elapsed = time_parser(parser,
file_buf,nread,doHash,
encoding,
NUM_TEST_RUNS);
fprintf(stderr,"Document parsed (%s, hash=%s, filterSpaces=%s): %ld usec\n",
parser_str,
doHash?"true":"false",
doFilterSpaces?"true":"false",
usec_elapsed);
}
fprintf(stderr, "Done\n");
return 0;
}
int main(int argc, char **argv) {
if (argc < 3) {
print_usage(argv[0]);
return 1;
}
if (strcmp(argv[1], "--h") == 0 || strcmp(argv[1], "--help") == 0 ) {
print_usage(argv[0]);
return 1;
}
g_log.m_disabled = true;
// initialize library
g_mem.init();
hashinit();
// current dir
char path[PATH_MAX];
realpath(argv[1], path);
size_t pathLen = strlen(path);
if (path[pathLen] != '/') {
strcat(path, "/");
}
g_hostdb.init(-1, NULL, false, false, path);
g_conf.init(path);
ucInit();
// initialize rdbs
g_loop.init();
g_collectiondb.loadAllCollRecs();
g_statsdb.init();
g_posdb.init();
g_titledb.init();
g_tagdb.init();
g_spiderdb.init();
g_doledb.init();
g_spiderCache.init();
g_clusterdb.init();
g_linkdb.init();
g_collectiondb.addRdbBaseToAllRdbsForEachCollRec();
g_log.m_disabled = false;
g_log.m_logPrefix = false;
uint64_t docId = strtoul(argv[2], NULL, 10);
logf(LOG_TRACE, "Getting titlerec for docId=%" PRIu64, docId);
Msg5 msg5;
RdbList list;
key96_t startKey = Titledb::makeFirstKey(docId);
key96_t endKey = Titledb::makeLastKey(docId);
msg5.getList(RDB_TITLEDB, 0, &list, startKey, endKey, 500000000, true, 0, 0, -1, NULL, NULL, 0, true, NULL, 0, -1, -1LL, false, true);
if (list.getNumRecs() != 1) {
logf(LOG_TRACE, "Unable to find titlerec for docId=%" PRIu64, docId);
cleanup();
exit(1);
}
XmlDoc xmlDoc;
if (!xmlDoc.set2(list.getCurrentRec(), list.getCurrentRecSize(), "main", NULL, 0)) {
logf(LOG_TRACE, "Unable to set XmlDoc for docId=%" PRIu64, docId);
cleanup();
exit(1);
}
logf(LOG_TRACE, "XmlDoc info");
logf(LOG_TRACE, "\tfirstUrl : %.*s", xmlDoc.size_firstUrl, xmlDoc.ptr_firstUrl);
logf(LOG_TRACE, "\tredirUrl : %.*s", xmlDoc.size_redirUrl, xmlDoc.ptr_redirUrl);
logf(LOG_TRACE, "\trootTitle : %.*s", xmlDoc.size_rootTitleBuf, xmlDoc.ptr_rootTitleBuf);
// logf(LOG_TRACE, "\timageData :");
logf(LOG_TRACE, "\t");
loghex(LOG_TRACE, xmlDoc.ptr_utf8Content, xmlDoc.size_utf8Content, "\tutf8Content:");
logf(LOG_TRACE, "\tsite : %.*s", xmlDoc.size_site, xmlDoc.ptr_site);
logf(LOG_TRACE, "\tlinkInfo");
LinkInfo* linkInfo = xmlDoc.getLinkInfo1();
logf(LOG_TRACE, "\t\tm_numGoodInlinks : %d", linkInfo->m_numGoodInlinks);
logf(LOG_TRACE, "\t\tm_numInlinksInternal : %d", linkInfo->m_numInlinksInternal);
logf(LOG_TRACE, "\t\tm_numStoredInlinks : %d", linkInfo->m_numStoredInlinks);
int i = 0;
for (Inlink *inlink = linkInfo->getNextInlink(NULL); inlink; inlink = linkInfo->getNextInlink(inlink)) {
logf(LOG_TRACE, "\t\tinlink #%d", i++);
logf(LOG_TRACE, "\t\t\tdocId : %" PRIu64, inlink->m_docId);
logf(LOG_TRACE, "\t\t\turl : %s", inlink->getUrl());
logf(LOG_TRACE, "\t\t\tlinktext : %s", inlink->getLinkText());
logf(LOG_TRACE, "\t\t\tcountry : %s", getCountryCode(inlink->m_country));
logf(LOG_TRACE, "\t\t\tlanguage : %s", getLanguageAbbr(inlink->m_language));
}
loghex(LOG_TRACE, xmlDoc.ptr_linkdbData, xmlDoc.size_linkdbData, "\tlinkdbData");
logf(LOG_TRACE, "\ttagRec");
TagRec *tagRec = xmlDoc.getTagRec();
for (Tag *tag = tagRec->getFirstTag(); tag; tag = tagRec->getNextTag(tag)) {
SafeBuf sb;
tag->printDataToBuf(&sb);
logf(LOG_TRACE, "\t\t%-12s: %s", getTagStrFromType(tag->m_type), sb.getBufStart());
}
logf(LOG_TRACE, "\t");
logf(LOG_TRACE, "Links info");
g_log.m_disabled = true;
Links *links = xmlDoc.getLinks();
g_log.m_disabled = false;
for (int i = 0; i < links->getNumLinks(); ++i) {
logf(LOG_TRACE, "\tlink : %.*s", links->getLinkLen(i), links->getLinkPtr(i));
}
cleanup();
return 0;
}
